Studies on chloramphenicol acetyltransferases specified by R-factors.

(1) Chloramphenicol acetyltransferases, specified by R-factors, were studied in Escherichia coli. Strains were constructed which contained two compatible R-factors, one F-like and one non-F-like, coding respectively for a type I and a type III variant of chloramphenicol acetyltransferase. In the majority of dual plasmid strains studied, the presence of the type I variant only was detectable. A strain carrying plasmids R429 (fi+) and R387 (fi-) synthesised both variants of chloramphenicol acetyltransferase. Possible mechanisms leading to the apparent dominance, in most dual plasmid strains, of the synthesis of the type I variant are discussed. (2) A mutant of E.coli J53(R429) was obtained which exhibited a four to five-fold increase in the production of chloramphenicol acetyltransferase. This mutant was characterised with respect to the effects of nutrient supply and growth rate on the production of chloramphenicol acetyltransferase, with a view to obtaining the maximum yield of enzyme from a four hundred-litre preparation of the organism. (3) The primary structure of two variants of chloramphenicol acetyltransferase (those specified by plasmids R429 and R387) were studied. The aim was to establish the extent of sequence homology exhibited, by each, with a prototype variant (type I) of known amino acid sequence. The R429-specified variant (type I) was seen to be highly homologous with the prototype structure while the type III variant demonstrated regions of greater sequence variability. The significance of these findings are discussed in relation to the known function, in the native enzyme, of certain regions of the prototype sequence. (4) The subunit interactions of chloramphenicol acetyltransferase were studied using chemical modification techniques. Extensive modification of the tetramer with acid anhydrides did not promote the dissociation of the enzyme into its component subunits. The quaternary structure was also relatively stable to extremes of pH. Extensive modification by methyl acetimidate produced a tetrameric protein of reduced catalytic activity. Preparative amidination experiments enabled the identification of those lysine residues which are inaccessible to modification in the native enzyme. No sequence homology was apparent between the unique 'buried' lysine-containing peptides from types I and III variants of chloramphenicol acetyltransferase. (5) The symmetry of chloramphenicol acetyltransferase was investigated by crosslinking with bifunctional reagents. The maximum yield of crosslinked products was poor (23-30%) and conclusions about the symmetry of the oligomer were tentative. The results indicate a possible 'dimer of dimers' arrangement of subunits in the R429-specified variant. (6) Tetrameric, hybrid chloramphenicol acetyltransferases, composed of subunits of the two variants specified by plasmids R429 and R387, were isolated from a dual plasmid strain of E.coli. An analytical investigation was made of the amino groups which were inaccessible, in the native enzyme, to modification by methyl acetimidate. The 'buried' lysines of the a3B a2B2 hybrids were observed to be different from those expected on the basis of a combination of a (R387-Specified) and B (R429- specified) subunits as seen in the parental (a4, B4) enzymes. This variation is tentative evidence for the participation of these lysine groups in interactions at the subunit contact regions of the tetramer. Based on this evidence, models for the subunit interactions of chloramphenicol acetyltransferase are proposed.